They will only further undermine their judicial and leadership credibility. The correct action would be to ack the usefulness of p2p, not only for distribution of the popular, but also as a giant communications and learning tool for all of mankind. We should be harnessing such a tool and using it for the greater good, not condemming it due to a few dinosaurs. Such a move (further?) ties government and the entire political process to the shit-eating lobbyists and others who blindly refute the pervasiveness and efficacity of p2p networks.
P2P does a better job than their thing based distribution networks, *and that isn't going to change* regardless of legislation. In fact, it's only going to get better (underground networked bittorrent servers, i2p networks, tunneled dynamic trust based distributed file and database server networks, etc). End result? final capitulation of government to the giant inept or perhaps, if we're lucky, public knowledge of this nonsense followed by mass boycott of those involved, but most likely, such actions will result in more and more public money funneled to public-good actions, with a good amount of that going right into the pockets of those who demonize the right way of doing things. Look at the war on X where X=[drugs,terrorism,etc]. Those getting rich are the ones who are spreading fear to the masses. Drumming up support for private interest under the poison mask of public good. This can't last much longer... surely we're smarter than that.
I pay more now for media than ever before (and am getting more from it than ever before, but only due to good homebrew filter techniques). Perhaps it is our lot to filter out the worthy 5% from the rest of the dregs, be it on broadcast networks or in politics. It's too bad the percentages aren't higher and the will to move in the right direction takes a back seat to filling the pockets of the few.
They've been blocking 80 for non business customers for years. Since the Cadvision takeover (2000?) anyway. Residential: $36.95, Business: $89.95. The future is monopoly.
As an aside, the bandwidth for the University of Calgary is provided by Telus. The negotiated contract was for campus wide %99.999 uptime and 10ms redundant failover. As a result, bandwidth is super expensive for the campus, with bit-mongering galore. Students who have paid large sums of money for their education are sitting in a communications vacuume while on or thru U of C facilities. There seems to be very little knowledge of the situation outside campus networking and no will or momentum (or ability?) to change the system and allow for more thruput and more research/community based higher-bandwidth usage.
When AGT (a public entity) was privatized it became Telus. Since the privatization the service level has not increased much (not even keeping up with tech advances). There's more advertising and more PR and more expansionism, but little more in the way of services. Unfortunatly for western Canada, there's very little other choice.
Yeah, it's called multicast, a router level ip broadcast protocol. It was shot down by (most of) the isp's because of the belief that it would lead to more bandwidth usage and because of the (slightly) higher hardware costs (and costs of upgrading hardware). It's the right way to do it that was not persued due to financial reasons. Now, they have to deal with duplication of unicast streams which use up much, much, much, much more bandwidth.
As an eclipse mirror, I did nearly 3TB of eclipse uploads a couple months ago. P2P (bittorrent, etc) would certainly reduce that traffic, but the eclipse mirror system doesn't seem to worry so much about bit counting. If I were to do similar in downloads of movies, I'd get most of the good stuff out there.
Trying 127.0.0.1...^MConnected to localhost.^MEscape character is '^]'.^MWelcome to MLdonkey^MWelcome on mldonkey command-line^MUse ? for help^MMLdonkey command-line:
> s futility^MResult of search 1^MReinitialising download selectors^M88 results (-152 waiting)
[ Num ] Names
[ 1] The War on Drugs - Ripped By Falken 785Mo.avi
[ 2] MGM and the end of rationality. Newworld. avi
[ 3] All little Timmy's seratonin are belong to us.wmv
[ 4] Pssst. You. That little closet projects of yours of the last few years is better than what we could cook up in decades, so now it is illegal.pdf
[ 5] An office clerks guide to open heart surgery.lit
[ 6] Misunderstanding distributed chaotic systems.pdf
> d 4 2
download started
download started
> q
Connection closed by foreign host.
Quad gpu is 8 video outs, which is enough for 2 overlapping stereoscopic projected 4 pane views into virtual worlds. Think 6'x5' rear projected panes each on the up, forward, left, and right. Then another set of projected screens overlapping those. One set of 4 panes is horizontally polarized, the other vertical and the operator wears polarized glasses. It would be a near surround 3d window into a real-time 3d environment. Admittedly, it will be a while 'til such installs are mass accepted (if ever), but the possibility is there.
Or, perhaps a home brew UUUXGA projected screen of multiple components for massive hd screenings. Many more possibilities, I'm sure. 2 computers firewire sync'd each with 8 outputs at 1600x1200 is a wall at 6400x4800. Assuming no artifacting, that would be very realistic.
ATI's card scaling tech seems to make possible up to 32 cards in a machine. Such a thing could immerse an entire audience in a 360 surround visually immersive 3d bubble.
Nokia is giving patents (and relying on patent system), but should linux kernel team accept patented stuff to begin with? I mean, it's nice and all for Nokia to share, but shouldn't they be sharing equally regardless? The patent system is notoriously slow, imprecise and contains many erroneous/broad patents. I can't help but think that relying on such a system is like willingly introducing a bottleneck. The whole pantent structure is also resting upon enforcement, and the assumption that people cannot cooperate in the first place. Both consumers and producers must realize that there is more momentum and vitality thru cooperation than thru regulated/enforced proprietarianism... I'd stay away from the patented stuff to begin with. If they are well intentioned towards openness, let them go the whole way.
Yes, well, it's perhaps a philosophical viewpoint more than practical. Working with objects from an object language is nice (and 'natural'). There is also the promise of versioning, quick object traversal and simple querying. In practice, those are very hard to implement seemlessly and standards for which are few (odmg, jdo). I've been using a homebrew object-relational mapping layer for years. It's slow (needs optimization), cumbersome to maintain (somewhat less than others, needs use optimization) and quite proprietary... but, it allows for dealing with (data) objects at a high level... modelling from the objects to the tables and not the tables to the objects. With a couple months work, the thing could be quite nice, replacing a 'real' (read expensive) oodbms (for smaller scale apps), but with the x-app sql binding and less data layer lock-in (and ability to migrate to more capable rdbms backends). I'd like to get it to the point where the changes to the object model roll across to the dbms schema (with override), which would allow for a much more fluid relation between code and data... and half the fun is doing it.
The fact that RDBMS' have been around for years means they've been under the scrutiny of many eyes for many years and are rock solid, if cumbersome. Of course, the fact that the few existing high performance oodbms' that do exist are horrifically expensive. Objectivity being the best I've seen... able to capture objects at many Gigabytes/Sec into multinode distributed object network (for CERN particle data), but, yow, oracle is cheap in comparison.
Podcasting is somewhat interesting.. simple, but polled... no different than rss, really. Major problem I found with it is the large hit on the server... a large file requires n times the size to download (where n is number of listeners).
Torrent pod casting requires a more complex client, but eliminates the problem. Torrent is downloaded, then download of torrented (large media) file starts, distributing the download over the network. A torrent casting video podcast would be a thing of beauty, especially when paired up to a collaborative media metadata backend (ie dyn website). RSS torrents are already supported thru azureus, as is i2p anon transmission layer... tho configuring such a thing is tricky. There is much progression in the torrent space, all of which applies to subscribable torrents.
A dedicated app (perhaps even re-packaging the azureus libs, large tho they are) would be quite useful, reducing the tech barriers. Further use optimization would be nice (ie the 'copy and paste rss url' must go... replaced with click to subscribe or, a standard selection of rss's from rss'd list of rss's which can be managed in app).
Podcasts are interesting, but limited, don't rule out something based on the technology, tho, especially if made easier to use.
G4's E3 coverage has been pretty good. Nice, somewhat fluffy, coverage of the new 3 yesterday was quite good. They don't dive too deeply into the hardware, and tend to get caught up on the little details, but better than nothing. Personally, I'm waiting for the end of the sea of fps and mondo destruction time wasters... it's got a f*cking teraflop, use it for benefit.
Hmm, yeah, I've been told of the clumsyness of xslt on several occasions, but, in practice it works fine for me.
My web engine (built and used over the last several years) fetches objects from object-relational mapping layer, gathers and performs operations in servlet, which results in collection of objects, then is marshalled to xml, transformed with xsl resulting in html (currently). The relational-object step allows for (friendly) data validation, model coherance, and high level coding. The xsl templates are hooked with FAM, so, when the underlying xsl file changes, the old templates are dumped and regenerated, which allows for work on pages in real-time. (OR/OXML mapping is not yet real-time, requiring restart on model or mapping change). I've built up xslt libraries for common object->html rendering. All together, it allows for some pretty quick coding, tho with a steep learning curve, I'd imagine.
XSLT does have it's faults, however. One of the premises is to provide a simple dom query language that can be used by interface designers, loosely binding to the work of the backend coders. In practice, xslt is too complex for those with little procedural language skills, and, as I work alone, have few beefs with the other half of the team. Perhaps I should take a look at the dom manipulation within laszlo, if only to explore the options within the realm. It would also be nice to break the submit-process-result cycle required by http... tho it can be worked around by using AJAX. XForms will also smooth things, possibly making the xml round-trip, with changing data always living in a synced dom. I'm loathed to put too much processing in the presentation step, however. One nexus of complexity is enough.
Thanks for the link. Looks interesting. XSLT transform on xml object data for dynamic laszlo sheets would be funky. Funkier still when laszlo speaks SVG for full browser, non-plugin rendering action. Interesting.
It's been a while since I used JBuilder, and use Eclipse (integration builds) daily. JBuilder came from the knowledge in Delphi, ie, good fast gui builder. Eclipse is not a gui builder (out of the box)... there are plugins which provide this functionality. The JBuilder I remember wasn't bad if you used it as intended. A bit of a quirky interface. Eclipse has such depth. From custom code formatting to macros to navigation to searching, etc. The openness of Eclipse is what makes it better than JBuilder. Find a bug? Hit up bugs.eclipse.org and it's fixed in the next release. The number of plugins is huge and the architecture of the system seems quite well done facilitating customization thru plugin building. Erm, and multi-lingual, and multi-formatting, and in-app web browsing and... Really, it's in a different class from JBuilder, and makes a wonderful wrench (after a somewhat tougher learning curve). A while ago, I realized what a gem Eclipse was and offered to become a download mirror... my coloc box is now pushing 4TB/mo of Eclipse builds!
I've a project in the line for archiving 10's of TB of data. Current solutions for accessing such a quantity of data are expensive and often without migration path. Blu-Ray/HD-DVD with >50G/disk brings all the benefits of higher data density. Less robotics, less media, quicker access, etc. In addition, other issues such as data permanance, thruput and migration paths can be rolled into the new technology. New media is important, and higher density media will become more and more important, especially with HD content rolling out, DVD-A, SACD, multi-channel video and continued digitization and archiving of existing analog data. I see multi-terabyte storage solutions becoming commonplace within 5 years, and those those currently large archives will get even bigger. Software and platforms for management of such a quantity of data are evolving and work hand-in-hand with larger density storage. It's quite important, and I'm glad to see the two main research groups pooling their talent.
This could be done with an 'active' installer. Installation profiles could be defined via a webapp, configured on server side. Profiles downloaded by installer, which contains core firefox code. As part of the process, installer prompts for level of install (Normal, WebDev, PowerUser, Custom, etc), extensions are then dl'd from repository and installed (or flagged to be installed on startup). I could see this being used in a windows environment, and wouldn't break manual install (gentoo, etc). Upkeep would be minimal, perhaps auto installing the latest version of the requested extension. Profiles would have to be limited to verified extensions, I guess.
However, Firefox is an excellent young browser, and one of its strengths is the minimal footprint. Perhaps such an enhancement would be a violation of the (successful) minimalist approach. Also, I don't think it is too much to ask of the user to to explore other aspects of a project on their own.
Of course, there are divx & dvd standalone players... some even do divx off dvd-r, which ~8hrs of dvd video, orso. Perhaps not as high fidelity as straight mpeg2, and only divx3 but perfectly fine for most uses (recorded tv, home video, etc).
DivX Certified Hardware and DivX Licensed Software
InterVideo WinDVR and WinDVD Creator Software Included
USB 2.0 Interface for Best Quality Video
Burn Direct-to-Disc and Edit-on-Disc Supported
One-year full warranty (parts, labor or replacement)
and, active linux support... way to go Plextor! OGG/Theora support would be a plus, but that's not stable, yet... still, use that upgradable firmware for something! Nice device. I'm planning on getting one.
Slashdot Mirror
P2P does a better job than their thing based distribution networks, *and that isn't going to change* regardless of legislation. In fact, it's only going to get better (underground networked bittorrent servers, i2p networks, tunneled dynamic trust based distributed file and database server networks, etc). End result? final capitulation of government to the giant inept or perhaps, if we're lucky, public knowledge of this nonsense followed by mass boycott of those involved, but most likely, such actions will result in more and more public money funneled to public-good actions, with a good amount of that going right into the pockets of those who demonize the right way of doing things. Look at the war on X where X=[drugs,terrorism,etc]. Those getting rich are the ones who are spreading fear to the masses. Drumming up support for private interest under the poison mask of public good. This can't last much longer... surely we're smarter than that.
I pay more now for media than ever before (and am getting more from it than ever before, but only due to good homebrew filter techniques). Perhaps it is our lot to filter out the worthy 5% from the rest of the dregs, be it on broadcast networks or in politics. It's too bad the percentages aren't higher and the will to move in the right direction takes a back seat to filling the pockets of the few.
1920x1200, hdtv spec, dvi/vga/composite/component in, PiP, etc $795. Group buy, in volume. Would make for great core of homebrew projector. Buy 10 of 'em and have a nice cave ;)
voices-for-change.com via nyud.net.
They've been blocking 80 for non business customers for years. Since the Cadvision takeover (2000?) anyway. Residential: $36.95, Business: $89.95. The future is monopoly.
When AGT (a public entity) was privatized it became Telus. Since the privatization the service level has not increased much (not even keeping up with tech advances). There's more advertising and more PR and more expansionism, but little more in the way of services. Unfortunatly for western Canada, there's very little other choice.
Yeah, it's called multicast, a router level ip broadcast protocol. It was shot down by (most of) the isp's because of the belief that it would lead to more bandwidth usage and because of the (slightly) higher hardware costs (and costs of upgrading hardware). It's the right way to do it that was not persued due to financial reasons. Now, they have to deal with duplication of unicast streams which use up much, much, much, much more bandwidth.
As an eclipse mirror, I did nearly 3TB of eclipse uploads a couple months ago. P2P (bittorrent, etc) would certainly reduce that traffic, but the eclipse mirror system doesn't seem to worry so much about bit counting. If I were to do similar in downloads of movies, I'd get most of the good stuff out there.
Trying 127.0.0.1...^MConnected to localhost.^MEscape character is '^]'.^MWelcome to MLdonkey^MWelcome on mldonkey command-line^MUse ? for help^MMLdonkey command-line: > s futility^MResult of search 1^MReinitialising download selectors^M88 results (-152 waiting) [ Num ] Names [ 1] The War on Drugs - Ripped By Falken 785Mo.avi [ 2] MGM and the end of rationality. Newworld. avi [ 3] All little Timmy's seratonin are belong to us.wmv [ 4] Pssst. You. That little closet projects of yours of the last few years is better than what we could cook up in decades, so now it is illegal.pdf [ 5] An office clerks guide to open heart surgery.lit [ 6] Misunderstanding distributed chaotic systems.pdf > d 4 2 download started download started > q Connection closed by foreign host.
Yeah, erm, I think the gigabyte solution will be somewhat under $200k.
Or, perhaps a home brew UUUXGA projected screen of multiple components for massive hd screenings. Many more possibilities, I'm sure. 2 computers firewire sync'd each with 8 outputs at 1600x1200 is a wall at 6400x4800. Assuming no artifacting, that would be very realistic.
ATI's card scaling tech seems to make possible up to 32 cards in a machine. Such a thing could immerse an entire audience in a 360 surround visually immersive 3d bubble.
Hmmm, the movie at ice.gov is quite large. Serving the file with bittorrent may help 'em out.
Nokia is giving patents (and relying on patent system), but should linux kernel team accept patented stuff to begin with? I mean, it's nice and all for Nokia to share, but shouldn't they be sharing equally regardless? The patent system is notoriously slow, imprecise and contains many erroneous/broad patents. I can't help but think that relying on such a system is like willingly introducing a bottleneck. The whole pantent structure is also resting upon enforcement, and the assumption that people cannot cooperate in the first place. Both consumers and producers must realize that there is more momentum and vitality thru cooperation than thru regulated/enforced proprietarianism... I'd stay away from the patented stuff to begin with. If they are well intentioned towards openness, let them go the whole way.
The fact that RDBMS' have been around for years means they've been under the scrutiny of many eyes for many years and are rock solid, if cumbersome. Of course, the fact that the few existing high performance oodbms' that do exist are horrifically expensive. Objectivity being the best I've seen... able to capture objects at many Gigabytes/Sec into multinode distributed object network (for CERN particle data), but, yow, oracle is cheap in comparison.
Torrent pod casting requires a more complex client, but eliminates the problem. Torrent is downloaded, then download of torrented (large media) file starts, distributing the download over the network. A torrent casting video podcast would be a thing of beauty, especially when paired up to a collaborative media metadata backend (ie dyn website). RSS torrents are already supported thru azureus, as is i2p anon transmission layer... tho configuring such a thing is tricky. There is much progression in the torrent space, all of which applies to subscribable torrents.
A dedicated app (perhaps even re-packaging the azureus libs, large tho they are) would be quite useful, reducing the tech barriers. Further use optimization would be nice (ie the 'copy and paste rss url' must go... replaced with click to subscribe or, a standard selection of rss's from rss'd list of rss's which can be managed in app).
Podcasts are interesting, but limited, don't rule out something based on the technology, tho, especially if made easier to use.
G4's E3 coverage has been pretty good. Nice, somewhat fluffy, coverage of the new 3 yesterday was quite good. They don't dive too deeply into the hardware, and tend to get caught up on the little details, but better than nothing. Personally, I'm waiting for the end of the sea of fps and mondo destruction time wasters... it's got a f*cking teraflop, use it for benefit.
Tried Horde?
My web engine (built and used over the last several years) fetches objects from object-relational mapping layer, gathers and performs operations in servlet, which results in collection of objects, then is marshalled to xml, transformed with xsl resulting in html (currently). The relational-object step allows for (friendly) data validation, model coherance, and high level coding. The xsl templates are hooked with FAM, so, when the underlying xsl file changes, the old templates are dumped and regenerated, which allows for work on pages in real-time. (OR/OXML mapping is not yet real-time, requiring restart on model or mapping change). I've built up xslt libraries for common object->html rendering. All together, it allows for some pretty quick coding, tho with a steep learning curve, I'd imagine.
XSLT does have it's faults, however. One of the premises is to provide a simple dom query language that can be used by interface designers, loosely binding to the work of the backend coders. In practice, xslt is too complex for those with little procedural language skills, and, as I work alone, have few beefs with the other half of the team. Perhaps I should take a look at the dom manipulation within laszlo, if only to explore the options within the realm. It would also be nice to break the submit-process-result cycle required by http... tho it can be worked around by using AJAX. XForms will also smooth things, possibly making the xml round-trip, with changing data always living in a synced dom. I'm loathed to put too much processing in the presentation step, however. One nexus of complexity is enough.
Thanks for the link. Looks interesting. XSLT transform on xml object data for dynamic laszlo sheets would be funky. Funkier still when laszlo speaks SVG for full browser, non-plugin rendering action. Interesting.
The early builds were ass. They became much better over time.
It's been a while since I used JBuilder, and use Eclipse (integration builds) daily. JBuilder came from the knowledge in Delphi, ie, good fast gui builder. Eclipse is not a gui builder (out of the box)... there are plugins which provide this functionality. The JBuilder I remember wasn't bad if you used it as intended. A bit of a quirky interface. Eclipse has such depth. From custom code formatting to macros to navigation to searching, etc. The openness of Eclipse is what makes it better than JBuilder. Find a bug? Hit up bugs.eclipse.org and it's fixed in the next release. The number of plugins is huge and the architecture of the system seems quite well done facilitating customization thru plugin building. Erm, and multi-lingual, and multi-formatting, and in-app web browsing and ... Really, it's in a different class from JBuilder, and makes a wonderful wrench (after a somewhat tougher learning curve). A while ago, I realized what a gem Eclipse was and offered to become a download mirror... my coloc box is now pushing 4TB/mo of Eclipse builds!
I've a project in the line for archiving 10's of TB of data. Current solutions for accessing such a quantity of data are expensive and often without migration path. Blu-Ray/HD-DVD with >50G/disk brings all the benefits of higher data density. Less robotics, less media, quicker access, etc. In addition, other issues such as data permanance, thruput and migration paths can be rolled into the new technology. New media is important, and higher density media will become more and more important, especially with HD content rolling out, DVD-A, SACD, multi-channel video and continued digitization and archiving of existing analog data. I see multi-terabyte storage solutions becoming commonplace within 5 years, and those those currently large archives will get even bigger. Software and platforms for management of such a quantity of data are evolving and work hand-in-hand with larger density storage. It's quite important, and I'm glad to see the two main research groups pooling their talent.
However, Firefox is an excellent young browser, and one of its strengths is the minimal footprint. Perhaps such an enhancement would be a violation of the (successful) minimalist approach. Also, I don't think it is too much to ask of the user to to explore other aspects of a project on their own.
Lamina leds are the best I've found. They've just released a small round array that looks interesting. Low power, efficient, balanced and bright.
Of course, there are divx & dvd standalone players... some even do divx off dvd-r, which ~8hrs of dvd video, orso. Perhaps not as high fidelity as straight mpeg2, and only divx3 but perfectly fine for most uses (recorded tv, home video, etc).
- First Official DivX Certified PC PVR Product
- Hardware Encode to DivX, MPEG-4, MPEG-2/DVD and MPEG-1/VCD
- Watch, Pause and Record Live TV
- High-Quality TV Tuner Included
- Free Electronic Programming Guide (EPG)
- Schedule Recording For When You Are Away
- Composite Video, S-Video, Composite Audio, RF/Coaxial Inputs
- DivX Certified Hardware and DivX Licensed Software
- InterVideo WinDVR and WinDVD Creator Software Included
- USB 2.0 Interface for Best Quality Video
- Burn Direct-to-Disc and Edit-on-Disc Supported
- One-year full warranty (parts, labor or replacement)
and, active linux support... way to go Plextor! OGG/Theora support would be a plus, but that's not stable, yet... still, use that upgradable firmware for something! Nice device. I'm planning on getting one.